1. Field of the Invention
The present invention relates to an apparatus for automatically translating a language and a method thereof.
2. Description of the Related Art
In general, an automatic translating apparatus reads a source text written in a source language (for example, English) and carries out morphological analysis. Then, a dictionary for translation is accessed to obtain grammatical information such as the part of speech for each word and a corresponding translation word for the target language, followed by analysis of the tense, person, number, and the like. An internal structure of the source language as a result of the morphological analysis and rules prepared in advance for that source language. The internal structure of the source language is transformed into an internal structure of the target language using a transformation rule. In accordance with the transformed internal structure, a translated sentence in the target language is generated using the translation words obtained by the dictionary look-up process during the morphological analysis.
In order to obtain an appropriate translation of the target language, an appropriate selection of a translation word for the target language is important. The selection of a translate on word conventionally includes the problems set forth in the following .
For example, the English language has the following problem regarding the auxiliary verb. Differing from other parts of speech such as the verb and noun, the auxiliary verb varies in its translation according to the "circumstance" in which it is used. This circumstance includes conditions such as whether the sentence including that word is an affirmative sentence or a negative sentence, or whether it is a declarative sentence or an interrogative sentence. Because the translation varies according to its circumstance as in the case of an auxiliary verb, just a simple translation setting will have a great side effect, making it harder to obtain an appropriate translation. For example, the word "may" alone will be translated into "suru kamo shirehal" in Japanese. In contrast, in the case of "may not", the translation should be "shite wa ikenai" Therefore, an adequate result can not be obtained just by setting a translation for an auxiliary verb of the English language to translate into a target language, for example Japanese.
In order to solve such a problem, there is a device having the priority allocated in advance for a translation word. A plurality of translation words are displayed according to this priority at the time of translation, whereby the operator selects an appropriate one. An example of such a technique is disclosed in Japanese Patent Laying-Open No. 61-90272.
A known improvement to the above-described technique includes update of the priorities of the translation words based on a learning process of the result selected by the operator. Such a technique is disclosed in Japanese Patent Laying-Open No. 63-136265, for example.
However, if the priorities for translation words are allotted and the display for selection is carried out according to the priority, the order of the displayed translation word or the selected result may not be appropriate for some subjects to be translated.
For example, consider the English auxiliary verb "shall". When the sentence including "shall" is an affirmative and declarative sentence in the first person, the translation output is "SURU DARO", otherwise "SURU KOTO NI NATTEIRU". However, in the case of translating a document related to the patent field, for example, the auxiliary verb "shall" should be output as "subeki" in most cases. In contrast, it is better to output the above former translation in a general literature.
In a conventional device where a priority is fixedly allocated to each translation word, the operator had to specify an appropriate translation for substitution every time this auxiliary verb appears in a sentence. In a conventional device where the display order of a translation word is changed according to a learning process of selected results, the learned result could be reflected only after the selection process was carried out several times in the case of translating documents of particular field such as patent related documents. Furthermore, the learned result obtained during translation of particular documents such as patent related documents would be lost if translation of a document of another field is carried out or when the power of the device is turned off. When the translation of a document of the same particular field such as patent related document is to be carried out subsequently, the specification of an appropriate translation had to be newly carried out.